RESEARCH Open Access Conditions for effective smart ... · RESEARCH Open Access Conditions for...
Transcript of RESEARCH Open Access Conditions for effective smart ... · RESEARCH Open Access Conditions for...
Koper Smart Learning Environments 2014, 1:5http://www.slejournal.com/content/1/1/5
RESEARCH Open Access
Conditions for effective smart learningenvironmentsRob Koper
Correspondence: [email protected] Chair at Institute Board,Open University of the Netherlands(Valkenburgerweg 177), Heerlen(6419 AT), The Netherlands
©Lp
Abstract
Smart learning environments (SLEs) are defined in this paper as physicalenvironments that are enriched with digital, context-aware and adaptive devices, topromote better and faster learning. In order to identify the requirements for ‘betterand faster learning’, the idea of Human Learning Interfaces (HLI) is presented, i.e. theset of learning related interaction mechanisms that humans expose to the outsideworld that can be used to control, stimulate and facilitate their learning processes. Itis assumed that humans have and use these HLIs for all types of learning, and thatothers, such as parents, teachers, friends, and digital devices can interact with theinterface to help a person to learn something. Three basic HLIs are identified thatrepresent three distinct types of learning: learning to deal with new situations(identification), learning to behave in a social group (socialization) and learning bycreating something (creation). These three HLIs involve a change in cognitiverepresentations and behavior. Performance can be increased using the practice HLI,and meta-cognitive development is supported by the reflection HLI. This analysis ofHLIs is used to identify the conditions for the development of effective smart learningenvironments and a research agenda for SLEs.
Keywords: Human learning interfaces; Smart learning environments;Research agenda
IntroductionWhy should we develop smart(er) learning environments? For me, the answer is to
promote better and faster learning. Every step forward from existing learning environ-
ments towards these smarter learning environments is an improvement that is essen-
tial for further human, economical and cultural development. This is a big challenge.
In this paper I will provide a contribution towards this ambition by exploring the con-
ditions that should be met by smart learning environments (SLEs) in order to stimulate
better and faster learning. For this purpose I will introduce a new theoretical concept,
named Human Learning Interfaces (HLIs), that can facilitate the research and develop-
ment of SLEs. I define HLIs as the set of interaction mechanisms that humans expose
to the outside world, and that can be used to control, stimulate and facilitate their
learning processes. An interface is the communication facility that enables two or more
systems to communicate with each other, i.e. facilitate the creation and interpretation
of messages that elicit a (learning) response (see Griffin 2012, p.6). It is assumed that
humans have and use these HLIs for all types of learning, and also others, such as
2014 Koper; licensee Springer. This is an Open Access article distributed under the terms of the Creative Commons Attributionicense (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium,rovided the original work is properly credited.
Koper Smart Learning Environments 2014, 1:5 Page 2 of 17http://www.slejournal.com/content/1/1/5
parents, teachers, friends and digital devices can interact with the interface to help a
person to learn something. For instance in education HLIs could be seen as the inter-
faces between the teacher and the learner: the teacher uses the interfaces of the learners
to enable them to learn something. The interfaces describe functions for input and out-
put: appropriate input for the senses of the learner to stimulate learning, observing
(output) behaviors and providing (input) feedback and feedforward. HLIs are abstrac-
tions of very complex sensory, cognitive and behavioral processes and can be utilized
in any situation to learn something, also by the person himself, or by elements in the
environment the person is in, including digital devices.
In the following sections I will elaborate the idea of HLIs. First I will identify and de-
fine some core concepts, followed by the description of the HLIs. Then I will present
the definition of Smart Learning Environments, providing a strong research challenge
for the field. In the discussion and conclusion I will discuss some research issues
related to HLIs and its use.
Core concepts
First I will inspect three core concepts that are needed for HLIs: learning environments,
smart learning environments and learning.
Learning environmentThe term Learning Environment is used in different contexts with different meanings
(see e.g. Abualrub et al. 2013). In the glossary of education reform a good definition is
provided (“Glossary” 2014): “Learning environment refers to the diverse physical loca-
tions, contexts, and cultures in which students learn.” Examples are classrooms, work-
places, labs, museums, natural sites, means of transport, and home. Most learning
environments are deliberately arranged or adapted to stimulate learning towards some
learning objectives, e.g. by adding learning materials, tasks, tests, feedback and support.
A learning environment can be arranged more or less generic, i.e. supports a smaller or
wider set of learning activities and learning objectives. Especially in modern learning
theories, the importance of embedding learning activities within authentic environ-
ments is emphasized (Vygotsky 1978; Brown et al. 1989), i.e. by solving real world
problems in the context in which they most typically occur (Merrill 2002).
There is also a more restricted use of the term learning environment, namely as an
abbreviation of ‘digital’ or ‘virtual’ learning environments. Using the previous definition
digital learning environments are digitally represented locations, contexts and cultures
in which students learn. Examples are serious games and virtual classrooms. One of the
problems of a focus on digital learning environment is that it leaves implicit that a
digital environment is always embedded in a physical environment. When using a digit-
ally represented environment, you are still located in a physical environment, such as a
living room, that can influence the learning processes and the cognitive representation
of the learning environment, e.g. by influencing attention processes, state of arousal,
memory cueing, concentration, encoding and recall. So, when considering digital learn-
ing environments, they should always be perceived as the addition of digital devices to
the current physical environment of humans, influencing the total cognitive representa-
tion of their environment, and through this, their behavior and learning processes. The
Koper Smart Learning Environments 2014, 1:5 Page 3 of 17http://www.slejournal.com/content/1/1/5
term ‘digital device’ is used here to represent the combination of some computer-based
hardware, such as a smartphone, robot, smart watch, smart glasses, smart board and
computer, and some dedicated software that provides the concrete functionality for the
learning process. Digital device(s) can be more or less aware of the physical context in
which they are used. Context-aware devices (Gross and Specht 2001), typically have
sensors that enable them to ‘know’ something about the environment they are in, in-
cluding some measurable aspects of the user. This information can be used to enrich
or augment the physical environment and to provide guidance and feedback to the
user.
Learning environments can be defined as the set of physical and digital locations,
contexts and cultures in which students learn. Five typical cases of learning environ-
ments can be distinguished with respect to their relation to digital devices:
1. The zero case: there are no relevant physical or digital relevant stimuli in the
environment of a person. The cognitive representations of the person can be
formed rather independent of the outside world: thinking, dreaming, visualising
something based on memory and creativity processes. In this case there is an
internally stimulated representation of the learning environment.
2. The digital case: when the physical environment includes digital learning devices,
but does not provide relevant non-digital stimuli to the user. For instance in a quiet
study room when using a simulation program. The representation of the learning
environment can dominantly be influenced by the digital device(s), e.g. by presenting a
virtual reality world, a serious game, a virtual classroom or a (digital) book. The
cognitive representations that are stimulated by the digital device can result in
learning processes. In this case there is a digital stimulated representation of the
learning environment.
3. The embedded case: the physical environment provides relevant stimuli to the user
and the digital devices are adding, augmenting information to enrich the cognitive
representation. In this case there is a combined, partly digital, partly physical
stimulated representation of the learning environment.
4. The side-by-side case: the digital devices are added to a physical environment to
support additional learning functions such as information, support, tests and
feedback, but the digital devices are ignorant of the actual physical environment. All
information about the physical environment should be added to the device by the
user. For example when students are presented with tasks to execute in their physical
environment, but they need to input the results to the digital device themselves. In
this case the user’s representation of the learning environment is fragmented: the
physical parts and the digital parts.
5. The classical case: the physical environment provides relevant stimuli, and there are
no additional digital relevant signals. This is ‘old school’ situation where humans are
interacting and learning without the help of any digital device. In this case there is a
representation of the learning environment by the user that is stimulated by the
physical environment.
Note that in any of the representations mentioned in the cases, different people have
different representations because they are selective in attention, represent stimuli in the
Koper Smart Learning Environments 2014, 1:5 Page 4 of 17http://www.slejournal.com/content/1/1/5
context of interests and pre-knowledge, and add and delete information to fit a view
which is comfortable to them. Behavior and learning are the result of these individually
different representations and not directly the result of the stimuli provided. This is why
observation of the behavior, feedback, feedforward and adaptations of interventions are
important when you want someone to learn something, e.g. in a smart learning
environment.
Smart learning environmentsWhat are smart learning environments? The idea of smart learning environments fits
in the tradition of adding the adjective ‘smart’ to various existing phenomena, such as
smart phones, smart tv’s, smart boards, smart lights and smart cities in order to identify
a next step in its development or a new generation. So, from this perspective smart
learning environments could be seen as learning environments that are considerably
improved to promote better and faster learning. So what type of improvements makes
a learning environment smart? First of all, when looking at the previous section about
learning environments, I would say that only the case ‘embedded’ could be considered
smart. The zero case and the classical case do not use digital devices, and an implicit
aim of smart learning environments is to use ICT to improve the learning environ-
ments. The digital case and the side-by-side case represent most of the current
computer-based learning solutions such as serious games, tutorials, drill and practice
and tests. They require the user to concentrate on the digital device that is responsible
for (parts of ) the learning process and the device itself is ignorant of the physical
environment the user is in. So, for me, the first set of requirements for SLEs are
the following. an SLE is a learning environment in which:
1. one or more digital devices are added to the physical locations of the learner;
2. the digital devices are aware of the learners location, context and culture;
3. the digital devices add learning functions to the locations, context and culture, such
as the provision of (augmented) information, assessments, remote collaboration,
feedforward, feedback, etc.;
4. a digital devices are monitoring the progress of learners and provides appropriate
information to relevant stakeholders.
So, an SLE is context-aware and adaptive to the individual learner’s behavior. How-
ever, concentrating on these technical aspects does not automatically promote better
and faster learning. I would even state that it is a loss of time when SLEs are build
without having concrete notions about the improvement in learning one wants to
accomplish. In the next section I will look at learning itself.
Human learningHuman learning has been studied extensively in various disciplines such as psychology,
cognitive sciences, neurosciences, pedagogy and anthropology. Schunk (2012, p.3) de-
fines learning as follows: “Learning is an enduring change in behavior, or in the capacity
to behave in a given fashion, which results from practice or other forms of experience”.
This definition contains three criteria: learning involves change, learning endures over
Koper Smart Learning Environments 2014, 1:5 Page 5 of 17http://www.slejournal.com/content/1/1/5
time, and learning occurs through experience and not by means of heredity for instance.
In this section some of the highlights of learning research will be summarized that are
relevant for HLIs.
The behavioristic approach to learning has concentrated on learning processes that
change or strengthen the association between stimuli and behavioral responses. It con-
siders the intermediate human cognitive processes as a black box that can not be
known and is not relevant to know for the explanation of learning. The cognitive sci-
ences have criticised and extended this view by opening up the black box by not only
looking at stimuli and responses, but also at cognitive processes that mediate their rela-
tionship. Furthermore some learning processes could not be explained by behavioristic
theories, notably the work of Bandura et al. (1961, 1963, 1973) made clear that people
could learn from observing the behavior of others, even without reinforcement. Also he
elaborated the notion of self-regulation (Bandura 1991), the ability to take control over
your own learning objectives, learning activities, learning environment and assessment
of progress. The cognitive tradition, inspired by information processing technologies,
asks questions such as: which learning related functions and processes are executed in
the brain? Ohlsson (2011, p.29) states that the central elements of cognition are mental
representations, and that differences in mental representations explain differences in
behavior. Cognitive functions (perception, remembering, thinking, acting and learning)
are implemented by processes that create, utilize and revise representations. The pro-
cesses are leading to three streams of events that should be explained by a general cog-
nitive theory: subjective experience, actions and utterances. Ohlsson (p.44) defines
learning as a meta-function for cognitive change and he states that it is implausible that
cognitive change can be explained by a single learning mechanism. He makes a distinction
between monotonic and non-monotonic learning mechanisms. Monotonic mechanisms,
such as associations, are governed by fixed, unchangeable rules and are predictable. Non-
monotonic learning mechanisms are complex learning mechanisms that are intrinsically
hard to predict, such as weather forecasts. Non-monotonic learning mechanisms are cog-
nitive processes that enable humans to suppress their experience and override its impera-
tive for actions (p.21). An example are creative processes that are closely linked to
learning, because in order to create something new, someone must change their cognitive
representations.
Functions and processes are expected to run within a cognitive architecture. The most
widely used architecture is the modal model, an adapted version of the Atkinson and
Shiffrin model (1968, 1971). Figure 1 represents the modal model with the distinct
memory stores positioned between input and output.
One interesting perspective on cognition is the dual process theory (Evans and Over
1996; Sloman 1996; Stanovich 1999, Stanovich 2004; Kahneman and Frederick 2005),
there is evidence that humans utilize two (or more) different cognitive processes at the
same time, often referred to as system 1 and system 2, or type 1 and type 2. Type 1
processes are associative, fast, automatic, needs low effort, have a large processing
capacity and work on the basis of learned heuristics. Type 2 processes are rule-based,
slow, have limited capacity because of limitations of working memory, is consciously
controlled and needs a high level of mental effort (Evans 2009). The idea is that all
stimuli that are beyond sensory memory are always processed by type 1 processes, lead-
ing to an immediate intention for a response. Type 2 processes ‘controls’ the response
Figure 1 The modal model of the architecture of human information processing (adapted fromAshcraft and Radvansky 2010 p.38).
Koper Smart Learning Environments 2014, 1:5 Page 6 of 17http://www.slejournal.com/content/1/1/5
intentions and can rethink critically the assumed associative, often fake causal relation-
ships used by type 1 processes. Type 2 is reflective and can approve the proposed
action, cancel it, or change it. Because the activity of type 2 costs energy, people can be
too lazy, too miser or too depleted (Baumeister et al. 1998) to control type 1 processes.
Most of the actions of type 1, when appropriately practiced and learned could be very
effective and efficient, but especially studies of Kahneman and Tversky (e.g., 1974,
1979) have shown that in more complex situations, especially when probabilities should
be calculated or algorithms should be used, type 1 processes make the wrong guesses.
This dual process theory comes close to another tradition of cognitive research: meta-
cognition and especially self-regulation (e.g., Butler and Winne 1995). Metacognition
refers to knowing what one knows and self-regulation is the process to govern activities
in line with goals, environment and other priorities. A high level of self-regulation
capabilities is needed for autonomous learning.
Besides cognition as a generic process, an important sub-discipline of social psych-
ology should be mentioned: social-cognition, the field that is focussed on the relation-
ship between human social behavior and cognition, including social learning as has
been discussed earlier (see Carlston 2013). One of the concepts that are studied in so-
cial cognition is priming, a non-declarative long term memory function, the effect that
the reaction to a stimulus is influenced by another stimulus (Meyer and Schvaneveldt
1971). Higgins et al. (1977) showed that introducing trait concepts in one task influ-
ences the trait related information in a non-related subsequent task. Many different
kinds of priming have been found and can for instance be used to contextualise the
interpretation of concepts to be learned. Of interest are also studies of socialization,
i.e. the process in which individuals are learning the perceptions, norms, values and
actions that are necessary to participate in a social groups such as family, work, peer
groups, organisations. (e.g. Mayer 2013; Hafferty and Hafler 2011).
Koper Smart Learning Environments 2014, 1:5 Page 7 of 17http://www.slejournal.com/content/1/1/5
Human learning interfacesGeneric model of HLIs
Humans interface to the outside world through their senses (input) and their behavior
(output). When designing an SLE, the environment must provide the right inputs and
monitor the outputs in order to stimulate or facilitate learning processes. The use of
these inputs and outputs can be generalized in a Human Learning Interface. In its most
generic form it contains an intervention mechanism that provides the input for the
senses, observations to monitor the behavior and physical state of the person, and ob-
servations to monitor changes in the physical environment. To control the learning to-
wards certain learning objectives, it should be possible to specify a goal and to compare
the observations with the goal to redirect the interventions (Figure 2).
A smart learning environment is identical to the physical environment of the user,
with the addition of some digital devices that utilize the HLI (white box in Figure 2). In
some situations the physical environment of the user is already arranged as a learning
environment, such as a class, but this is not necessarily so. Digital devices can be added
to any environment to create a learning experience. In social learning the digital devices
also interact with other humans, e.g. to create a learning community. The core
elements of the HLI will now be discussed.
Interventions
What type of interventions can be distinguished? I make a distinction between four core
interventions for learning:
1. Ask a question to the learner. Asking questions invokes all kinds of cognitive
processes. It invokes priming (questions offer a context), retrieval of information, it
Figure 2 General overview of an human learning interface.
Koper Smart Learning Environments 2014, 1:5 Page 8 of 17http://www.slejournal.com/content/1/1/5
invokes higher order thinking and testing effects. With respect to testing effects,
research has shown that people can better be tested with questions, e.g. self tests,
than studying texts in the regular way using mind maps and notes (Carrier and
Pashler 1992; Karpicke and Blunt 2011).
2. Give the learner a task to accomplish. Tasks can be very basic (“Look over there, a
fox!”) to highly complex, such as problems to solve, decisions to make. They can
also be an instruction to read or study something (“Read this book”).
3. Provide something to the learner. The provisioning of information, an event that
occurs in the physical environment, some instruments or other resources that are
given to the learner. Provisioning of events, access to persons and tools are part of
the arrangement of the learning environment. A learning environment should have
the appropriate tools, information and people available in order to facilitate certain
learning processes.
4. Conditioning of the environment of the learner. Provide positive and negative
feedback, provide incentives, create contingencies, or provide associative stimuli. In
most learning environments this is provided by a teacher who praises students for
behavior, provide grades, or provide some other forms of incentives. In digital
learning environments you also see batches, points and rankings to condition the
environment.
The assumption is that more complex instructional interventions can be build from
selections and sequences of these core intervention mechanisms. For instance, assess-
ments can be build as combinations of information, tasks, questions and feedback. In
SLEs I expect these selections and sequences to be adaptive: the next interventions are
dependent on the results of the previous interventions.
Observations
Observations of a) relevant behaviors, including relevant physical changes and b) rele-
vant changes in the physical environment can be done by humans, or by sensors of
digital devices. The observations determine how adaptive an SLE can be. For each of
the interventions expected output behaviors should be specified to determine what
should be observed. A simple example: a digital device provides the task to look at a
painting in a museum and asks the question to identify what the main color pallet is
that is used in the painting. Possible behaviors of the learner could include to verbally
state a color pallet, to look a bit confused, not know what to answer, to ignore (parts)
of the instruction, or to look at something else. In order to provide the correct next
intervention, the sensors should be able to observe these states. Observations are also a
necessary condition for feedback and assessment. One of the challenges for SLEs is the
further development of useful sensors and its use to infer behaviors or changes in the
environment.
Objectives
Specific learning objectives can be set by a user, e.g., a learner or a teacher, of the HLI
to evaluate progress towards a defined learning outcome. It is not necessary to set
objectives in advance to make the HLI function, it could also consist of a series of
Koper Smart Learning Environments 2014, 1:5 Page 9 of 17http://www.slejournal.com/content/1/1/5
interventions only, including observations of behavior, but without control. In special
cases this could be an approach. Most cases of learning are directed towards certain
learning outcomes and need clear objectives to specify these outcomes. For example
the taxonomy of Bloom (Bloom et al. 1956; Krathwohl et al. 1964; Anderson et al.
2001) can be used for this purpose.
Different kinds of HLIsAs presented in the paragraph about human learning, different types of learning can be
distinguished and for each type I will now define a specific HLI. The categorization I
propose is based on the central concept of cognition: representations. What different
kinds of representations and related behaviors and learning processes can be distin-
guished? First of all I will make a distinction between representations of the physical
world and representations of the social environment as a subset of the physical world.
Representations and related actions of the physical world are needed for adaptation
and survival of the individual. “Individuals explore the world and dynamically acquire
representations and modify them as needed…” (Sun et al. 2009, p.243). Representations
of the affordances and constraints in the social world are needed to adapt to social
groups and to carry over cultural learned knowledge and skills between group members
and between generations. Members of a species have innate mechanisms to interact
with members of the same species and these are distinct from other cognitive pro-
cesses. For instance people have a very well developed attention mechanisms and mem-
ory for human faces and are very sensitive to facial expressions (Hugenberg and Wilson
2013). These two types of representations are related, because the social representations
always occur in a physical context. However, learning could be more directed towards
the physical world than the social world. Furthermore, in some cases humans can be
represented as objects, instead of subjects, e.g. when walking in a busy street. And also,
physical objects can be humanized as is done by kids and people who attribute underlying
intentions to natural phenomena. All elements of the modal model are expected to be
present in both situations, only the type of representation (physical or social) is different.
A next type of learning is based on construction, creation. People are able to create,
to design and construct artefacts and new behaviors in interaction with the facilities
that are provided in the world. As is stated before, these creative processes involve cog-
nitive re-representations of the world and, because of this, to learning. Another type of
learning is directly linked to the encoding and retrieval of representations of knowledge
and actions in long term memory. The more something is rehearsed, the more efficient,
automatic and faster it becomes. Rehearsal can be seen in two ways (Craik and Lockhart
1972): maintenance rehearsal in which you rehearse to prevent forgetting and elaborative
rehearsal to provide better connections with prior knowledge (schema’s) in long term
memory. This type of rehearsal improves the storage and recall of knowledge, but it
can also improve representations of actions in order to perform better and faster. I
will refer to this category as practice. Each of the previous three types of learning can
be practiced: learning to cope better and faster with physical situations, with social
situations and to be able to create something better and faster. The last category I
distinguish is the type 2 processing of the central executive in working memory: decision
making, planning, overruling experience, and self-regulation. This is about representations
of representations (metacognition) and critical reflection of internal representations and
Koper Smart Learning Environments 2014, 1:5 Page 10 of 17http://www.slejournal.com/content/1/1/5
behavioral impulses. Following this distinction in learning processes, five different HLIs
can be identified, each with a different learning goal (Figure 3):
1. Learning to represent new situations and events in the world and know how to act
and react. This is facilitated through the identification HLI.
2. Learning to represent the social norms, values, customs and ideologies of social
institutions and learning the skills and habits that enables you ‘to behave’ within the
social institutions, including the dissemination of norms and values to others. Social
institutions include: family, peer groups, religion, economic system, language and
legal system. social situations and events given. This is facilitated through the
socialization HLI.
3. Learning to represent new sequences of behavior in order to create something. This
is facilitated through the creation HLI.
4. Learning to represent knowledge and actions faster and better, including the
representation of performance targets or future incentives for any of the previously
mentioned behaviors. This is facilitated through the practice HLI.
5. Learning to create representations of representations and to change the initial
representations and behaviors. This is facilitated through the reflection HLI.
In the next paragraphs, each of these HLIs will be discussed.
The identification HLIIdentification is the learning mechanism that is needed when you encounter a new situ-
ation and you learn through experience and acting in the situation. It is like dropping
someone in an unknown territory that you have to explore and deal with. The basic
learning activities are: exploration, recognition, differentiation and generalization of
stimuli, labeling (groups of ) stimuli, building knowledge about the behaviors (dangers,
contingencies, rules, affordances) of the unknown stimuli, creating mental maps
Figure 3 The five HLIs.
Koper Smart Learning Environments 2014, 1:5 Page 11 of 17http://www.slejournal.com/content/1/1/5
(schema’s) of the environment, acting within the situation and solving problems.
Some typical learning interventions in this HLI are the following:
� Provide a (simulated or real) situation which is new to the person. The situation
includes events and contingencies that challenges the person to react in an
appropriate way.
� Give the (implicit or explicit) task to observe and explore the situation and its
elements, including possibilities to act and experiment.
� Give the (implicit or explicit) task to identify and categorize differences and
commonalities (specialisation and generalization) in a way that the constituents of
the situation can be labeled with symbols or names. Ask questions about this and
provide feedback.
� Provide stories about the situation, the constituents and its history to learn to know
the processes, dynamics, contingencies and characters. A story can be informal, but
also a scientifically justified theory. Ask questions about the stories and ask the
person to tell stories themselves and map the situation.
� Give the task to perform some activities within the environment, such as travelling,
using tools, driving, swimming, etc.
� Provide incentives when conditions are met.
� Give the task to solve some challenging problems and ask “what-if” questions.
Although the list above suggests an order or a pedagogical approach, this is not
intended. They should only be seen as possible examples of interventions in the Identi-
fication HLI. It is dependent on the concrete setting, the content to be learned, which
pedagogical model fits best. Note that the process can be social or nonsocial: exploring
new ground with a group or alone. Furthermore, in every step listed above, support can
be provided to the person in several ways: by adding a person (teacher; guide; parent;
expert) who points at important aspects of the environments and helps to label them
(“Look this is an eatable mushroom called champignon”) or a digital device with the
same functionality. In the context of SLEs this would typically be supported by a digital
device or by a combination of a digital device and a supporting person. In all cases the
best approach is to scaffold the support: slowly diminish the support to enable the per-
son to act independently (Lave and Wenger 1991).
To create SLEs that use the Identification HLI, it is important to identify prototypical
situations that represent the settings that the learner needs to cope with. For instance
typical situations that the person has to deal with in a (future) job.
The socialization HLISocialization is the process in which people acquire the cognitive, social and emotional
skills that are needed to function in the social community using the social skills for
which they are biologically prepared (Bugental 2000). It is for instance needed when
travelling to a new country, going to a new class or school, to a new job or having to
deal with new customers. Bugental and Grusec (2006) distinguish four types of
socialization relationships: attachment relationships that provide proximity between
members to maintain safety; social identity relationships to make people feel part of the
group and acquire group norms, values and skills; hierarchical relationships that facilitate
Koper Smart Learning Environments 2014, 1:5 Page 12 of 17http://www.slejournal.com/content/1/1/5
compliance, leadership and discourage maladaptive behaviors; and reciprocal relationships
that include the reciprocal exchange of knowledge, benefits and positive affect between
peers. They provided evidence that each of these relationships have a different socialization
task, developmental course, neurohormonal processes involved and evolutionary support.
Essential learning mechanisms in socialization are observational learning (Bandura 1977)
and imitation or mimicry (Chartrand and Van Baaren 2009). Examples of social behaviors
that are acquired by observation of others are: self-regulation, aggression, resistance to
temptation and moral judgement. People can participate peripherally in societies with the
intent to become full members later (Lave and Wenger 1991), during this peripheral par-
ticipation they are supported to learn the habits and skills of the group by observational
learning and supported action. Imitation seems ubiquitous, Dijksterhuis (2013, p.242)
states that research has shown that people mimic postures, gestures, facial expressions,
and a multitude of speech related phenomena. Interpretation of facial expressions play an
important role in socialization. Hugenberg and Wilson (2013, p.167) state that faces can
communicate a vast body of information quickly and efficiently, such as facial identities,
social categories (e.g., sex, race, age), physical health, evaluations, emotions, and intentions.
Facial expressions can provide immediate feedback about the mood of a person or group
and can be used to communicate appreciation or disapprovement as a reinforcer for
behavior. Some typical learning interventions in the Socialization HLI are the following:
� Provide a (simulated or real) situation in which a person encounters a new social
group, people with another cultural background or a variety of customers or people
to work with. The situation includes social events, demands and contingencies that
challenges the person to react in an appropriate way.
� Give the (implicit or explicit) task to observe the person or group and discover the
norms, values, habits and relationships.
� Give the tasks to try to meet and interact with group members, taking their values
and habits into account.
� Provide feedback to the person about the activities and answer questions.
� Give the task to mimic activities of others, especially skills and habits that are
typical for the group. For example in art groups (dance, painting, music) this is a
basic mechanism for learning.
� Give social tasks to execute, such as participating in a specific social and cultural
activities, selling a product to a customer, discuss an issue, celebrate something
collaboratively, etc.
� Provide incentives when conditions are met.
For SLEs socialization offers the challenge that not only direct communication
between actors should be supported, but also indirect communication that mediates
the socialization process, especially non-verbal behaviors, including facial expressions.
Also effective and non-intrusive sensoring of facial expressions, moods and gestures are
needed to fully support socialization in SLEs.
The creation HLICreation is the mechanism in which a person learns by manipulating the environment
in such a way that a new object, resource or process is made. What is produced can
Koper Smart Learning Environments 2014, 1:5 Page 13 of 17http://www.slejournal.com/content/1/1/5
have many forms, such as works of (applied) art, written texts, a meal, a decorated
room, a drawing, music, a house, etc. Creation processes involve cognitive change.
Creation and technological skills are closely related (Mitcham 1994). Some typical
learning interventions in this HLI are the following:
� Provide the (simulated or real) situation and materials that are needed for a
creation process.
� Give the task to a person to create something new and provide a framework for
incentives for good task accomplishment. The task can be performed alone or in a
group.
� Provide feedback to the person about the created artefact and answer questions.
� Give the task to a person to create something, given a set of requirements that
must be met, and provide procedures and worked-examples (Sweller 2006).
� Give the task to a person to restructure or decorate an environment.
� Ask questions to the person about the created artefact, e.g. its expression, its use,
its quality, how it could be improved.
� Provide incentives when conditions are met.
SLEs should enable to support learning before, during and after the creation process.
The creation process itself could take place in the physical world, such as creating a
painting on canvas, but can also be supported by a digital device such as a word pro-
cessor, a paint or design program or even a simulation or game in which you can build
structures and processes. In all cases the challenge is that the environment should contain
the physical and digital elements that enable the person to learn, including monitoring to
adapt interventions.
The practice HLIAll three of the previously mentioned HLIs, Identification, Socialization and Creation,
develop skills that could be further practiced to increase the performance on certain
tasks. The core characteristic of practice is to repeat activities to improve the tempo or
the quality of the result and to prepare for a high performance in future situations.
Practice of skills also leads to automation of these skills. Practice could be performed in
reality, but also in play. Play is a way of practice that can been seen in all mammals and
other species such as birds. Play has no sense in itself, but prepares the person to cope
with future real situations. People can play alone or with peers and represent mentally
(parts of ) the environment in which the activities are set (the playground), including
rules and performance criteria. Important aspects of practice and play are: a) there
should be a suitable (real or partly imaginary) environment, including tools and peers
for the activities to be performed, b) there should be the possibility to repeat the same
activity over and over again, c) performance measures and standards should be set,
along with incentives, d) feedback about the accomplishments per repetition should be
made available and e) incentives provided. Examples of learning environments that use
the practice HLI are classical drill and practice programs and serious games. But in fact,
all learning environments that support repetition and performance criteria are suitable
as a practice environment. In smart learning environments the practice devices are
expected to be context-aware and adaptive, e.g. like GPS sport watches could be used
Koper Smart Learning Environments 2014, 1:5 Page 14 of 17http://www.slejournal.com/content/1/1/5
as a digital feedback device to learn to perform better. The practice HLI is a meta-HLI
and envisaged to be integrated into each of the three previous HLIs.
The reflection HLIThe reflection HLI stimulation of system 2 (‘hard’) thinking: reflection, reframing of the
problem and solution, evaluation of results, decision making, strategy development,
self-regulation, and non-monotonic, deep learning as described by Ohlsson (2011). This
involves overruling previous experience, trying out new ideas, creative insight, adaptation
of cognitive skills by learning from errors, and conversion from one believe to another in-
compatible believe. Reflection can be stimulated as follows: a) stimulate people to take the
time to consider different options before they act; b) stimulate people in creative thinking,
coming up with new ideas and solutions, c) providing the possibility to review and analyse
their actions and effects and reflect on possible improvements, d) provide possibilities
to test yourself by asking questions, evaluate the objectives, environment and behavior,
e) stimulate the decision making processes, and facilitate strategy development. The
reflection HLI is also a meta HLI to all four previously mentioned HLIs. This means
that a smart learning environment that is focussed on Identification, this learning
environment should also include facilities for practice and for reflection in order to
stimulate faster and better learning. The same is true for Socialization and Creation.
Conditions for smart learning environmentsDefinition
Taking this all together, the concept of SLEs can now be defined as follows: SLEs are
physical environments that are improved to promote better and faster learning by
enriching the environment with context-aware and adaptive digital devices that,
together with the existing constituents of the physical environment, provide the situa-
tions, events, interventions and observations needed to stimulate a person to learn to
know and deal with situations (identification), to socialize with the group, to create
artefacts, and to practice and reflect.
It should be noted that the HLIs are defined independent of any specific content
domain, any specific pedagogical or instructional approach or any specific technologies in
mind. This makes the approach generic applicable as a theoretical model for all types of
smart learning environments. In practice of course, these specific choices should be made
before an SLE can be implemented and used. A second note is that I have focussed on the
core objective of SLEs, better and faster learning, but besides that core objective, SLEs also
need to be safe, reliable, inclusive and engaging. These latter aspects are not in scope
within this paper, but also need attention and further elaboration.
Conditions
To summarize the conditions for the development of effective SLEs:
1. An SLE is a learning environment in which one or more digital devices are added
to the physical locations of the learner to add learning functions, to add virtual or
augmented situations & locations, to add facilities for the monitoring and
assessment of the progress of learners and to add facilities for the provision of
appropriate information to relevant stakeholders;
Koper Smart Learning Environments 2014, 1:5 Page 15 of 17http://www.slejournal.com/content/1/1/5
2. With respect to the learning functions, the SLE supports at least one of the three
core HLI’s: identification, socialization and creation, depending on the objectives to
be attained with the SLE.
3. The SLE supports both meta HLI’s in order to stimulate practice and reflection on
the core HLI’s.
4. The SLE implements the functionality to observe the learners location, context,
preferences, physical and mental condition and culture and adapts interventions
based on relevant observations.
5. The SLE implements the intervention mechanisms: questions, tasks, provision of
information & resources, and conditioning.
6. The SLE implements functionality to specify and communicate learning objectives.
7. The processes in the SLE do not cause frictions, like waiting times and other
inefficiencies. The processes includes enrollment, teaching and learning,
examination and support and information facilities.
Discussion and conclusion: a research agenda for SLEsAs stated in the introduction, smart(er) learning environments promote better and
faster learning. In order to do so, I presented the idea of Human Learning Interfaces as
the set of interaction mechanisms that humans expose, and that can be used to control,
stimulate and facilitate their learning processes. I distinguished three basic HLIs
(identification, socialization and creation) and two meta HLIs (practice and reflection),
of which reflection represents higher-order thinking of type 2 as it is seen in dual
process theory of cognition. Based on this analysis the conditions for the development
of Smart Learning Environments are defined.
What future research is needed in order to develop SLEs that are utilizing the HLIs
effectively? I will now present a provisional research agenda in order to realize this
ambition of SLEs. Some of this research is more oriented at cognition, learning and
instruction and some more at the technological aspects of SLEs:
1) Which sequences of interventions stimulate faster and better learning in each of the
three core HLIs? The objective is to provide clear prescriptions how to arrange
interventions to help people to learn faster and better. The same question should be
asked for the meta-HLIs (practice and reflection). It could be expected that
the meta-HLI’s have some generic functions, but also specific functions per
core HLI. Furthermore, the HLIs should be more formalized to represent an
interface description that provides exact requirements for the developers of
SLEs, including a testing framework.
2) Which individual factors and differences are of importance for the adaptation of
interventions within SLEs? People differ on many dimensions and SLEs should
adapt to relevant differences. The first thing to consider are the differences in the
preferences of persons, i.e. preferences in ‘what’ to learn and ‘how’ to learn. For
instance, can these preferences, especially about ‘how to learn’, be summarized in
multidimensional preference profiles that are relevant for the selection of
interventions? Another aspect to take into account are differences in the physical
and mental condition, the motivation, concentration, self-efficacy and the personal
situation of the individual. Which of these differences are relevant triggers for
Koper Smart Learning Environments 2014, 1:5 Page 16 of 17http://www.slejournal.com/content/1/1/5
adaptations of interventions? It should be noted that individual differences research
is hard, because many attempts have failed to create impact (e.g. ATI research and
research on learning styles of Pashler et al. 2008). Adaptive systems have at the
moment very weak grounded effects on learning; more fundamental research into
individual differences and the adaptation of interventions towards these differences
is required in order to create more effective adaptive systems.
3) Which reactions of a person - as a consequence of an intervention - requires an
adaptation of the subsequent interventions and how are these reactions observed?
The observations can be formative or summative assessments of learning outcomes,
but also of responses to interventions: interventions do not always invoke the
expected reaction, e.g. people can be disturbed, have a different understanding or
provide a response that is inadequate. In a learning environment, most of these
observations are normally done by humans, e.g. teachers and peers. It would be
interesting to research whether (parts) of these observations can also be automated
by digital devices with adequate sensors. Part of this question is research which
adaptations of interventions are effective for each HLI based on observations of
individual responses, preferences, conditions, locations, context and culture? This is
a core question that can only be answered after answering the previous questions.
An important point to note is the term ‘needed’. Many adaptation rules can be
implemented, but are they really required to improve learning?
4) Which factors in the organization of the processes within the SLE (enrollment,
teaching and learning, examination, support and information) and which external
factors can disturb a smooth flow through the process for learners, teachers and
others, i.e. cause frictions, dissatisfaction and dropout? How can this be improved?
Examples are the inclusion of unnecessary waiting times, enrollment constraints,
grouping of students, inefficient temporal organization of the curriculum, the
degree of flexibility in tempo, etc.
5) How can we arrange SLEs that convincingly improve learning in comparison to
existing learning environments. Stated differently: how to develop SLEs or adapt
existing learning environments that meet the conditions for effective SLEs?
Experiments are needed, a catalogue of best practices, and also reference
architectures and other guidelines for the development of effective SLEs.
6) And last but not least, more fundamental research is also needed to the HLIs
themselves, for instance are they distinct categories at the neuro-cognitive level, or
mere interfaces that work on top of generic processes? Are there differences
between persons in the functioning of the HLIs? How do they develop over time?
Do the three core HLIs interact or are they distinct? Can they be trained?
Competing interestsThe author declares that he has no competing interests.
Received: 11 July 2014 Accepted: 22 September 2014
References
“Glossary”, The glossary of education reform, in Edglossary.org (Glossary of education reform). (2014). Retrieved July, 10, 2014,from http://edglossary.org/learning-environment/I Abualrub, B Karseth, B Stensaker, The various understandings of learning environment in higher education and its
quality implications. Qual. High. Ed 19(1), 90–110 (2013)
Koper Smart Learning Environments 2014, 1:5 Page 17 of 17http://www.slejournal.com/content/1/1/5
LW Anderson, DR Krathwohls, PW Airasian, KA Cruikshank, RE Mayer, PR Pintrich, J Raths, MC Wittrock, A Taxonomy forLearning, Teaching, and Assessing: A Revision of Bloom’s Taxonomy of Educational Objectives (complete ed.) (Longman,New York, 2001)
MH Ashcraft, GA Radvansky, Cognition, 5th edn. (Pearson, NJ, 2010)RC Atkinson, RM Shiffrin, Human memory: a proposed system and its control processes. Psychol. Learn. Motiv.
2, 89–195 (1968)RC Atkinson, RM Shiffrin, The control processes of short-term memory. Sci. Am. 225, 82–90 (1971)A Bandura, Social Learning Theory (Prentice Hall, Englewood Cliffs NJ, 1977)A Bandura, Social cognitive theory of self-regulation. Organ. Behav. Hum. Decis. Process. 50(2), 248–287 (1991)A Bandura, RW Jeffrey, Role of symbolic coding and rehearsal processes in observational learning. J. Pers. Soc. Psychol.
26(1), 122 (1973)A Bandura, D Ross, SA Ross, Transmission of aggression through the imitation of aggressive models. J. Abnorm. Soc.
Psychol. 63(3), 575–582 (1961)A Bandura, D Ross, SA Ross, Imitation of film-mediated aggressive models. J. Abnorm. Soc. Psychol. 66(1), 3–11 (1963)RF Baumeister, E Bratslavsky, M Muraven, DM Tice, Ego depletion: Is the active self a limited resource? J. Pers. Soc.
Psychol. 74(5), 1252–1265 (1998)BS Bloom, MD Engelhart, EJ Furst, WH Hill, DR Krathwohl, Taxonomy of Educational Objectives: The Classification of
Educational Goals (handbook I: cognitive domain) (David McKay Company, New York, 1956)JS Brown, A Collins, AP Duguid, Situated cognition and the culture of learning. Educ. Res. 18(1), 32–42 (1989)DB Bugental, Acquisition of the algorithms of social life: a domain-based approach. Psychol. Bull. 26, 187–209 (2000)DB Bugental, JE Grusec, Socialization processes, in Handbook of Child Psychology: vol. 3. Social, Emotional, And
Personality Development, ed. by W Damon, N Eisenberg, 5th edn. (Wiley, New York, 2006), pp. 366–428DL Butler, PH Winne, Feedback and self-regulated learning: a theoretical synthesis. Rev. Ed. Res. 65(3), 245–281 (1995)DE Carlston, The Oxford Handbook of Social Cognition (Oxford Press, Oxford, 2013)M Carrier, H Pashler, The influence of retrieval on retention. Mem. Cognit. 20, 632–642 (1992)TL Chartrand, R Van Baaren, Human mimicry. Adv. Exp. Soc. Psycol. 41, 219–274 (2009)FI Craik, RS Lockhart, Levels of processing: a framework for memory research. J. Verbal. Learn. Verbal. Behav.
11(6), 671–684 (1972)A Dijksterhuis, Automaticity. Handbook of social cognition, 2013, pp. 239–256JSBT Evans, How many dual-process theories do we need? One, two or many? in Two Minds: Dual Processes and
Beyond, ed. by JSBT Evans, K Frankish (Oxford Un. Press, Oxford, 2009), pp. 33–54JSBT Evans, DE Over, Rationality and Reasoning (Psychology Press, Hove, 1996)EA Griffin, A First Look at Communication Theory, 8th edn. (McGraw-Hill, Boston, 2012)T Gross, M Specht, Awareness in context-aware information systems. Mensch & Computer 2001 Berichte des German
Chapter of the ACM 55, 173–182 (2001)FW Hafferty, JP Hafler, The hidden curriculum, structural disconnects, and the socialization of new professionals. Innov.
Change. Prof. Educ. 6, 17–35 (2011)E Higgins, WS Rholes, CR Jones, Category accessibility and impression formation. J. Exp. Soc. Psychol. 13(2), 141–154 (1977)K Hugenberg, JP Wilson, Faces are central to social cognition, in Handbook of Social Cognition (2013), pp. 167–193D Kahneman, S Frederick, A model of heuristic judgment, in The Cambridge Handbook of Thinking and Reasoning, ed.
by K Holyoak, RG Morrison (Cambridge University Press, Cambridge, 2005), pp. 267–294D Kahneman, A Tversky, Prospect theory: an analysis of decision under risk. Econometrica. J. Econometric. Soc.
47(2), 263–291 (1979)JD Karpicke, JR Blunt, Retrieval practice produces more learning than elaborative studying with concept mapping.
Science 331, 772–775 (2011)DR Krathwohl, BS Bloom, BB Masia, Taxonomy of Educational Objectives: The Classification of Educational Goals. Handbook
II: the Affective Domain (David McKay Company, New York, 1964)J Lave, E Wenger, Situated Learning: Legitimate Peripheral Participation (Cambridge University Press, Cambridge, 1991)P Mayer, Socialization: the Approach from Social Anthropology (first print 1970) (Routledge, London, 2013)MD Merrill, First principles of instruction. Educ. Technology. Res. Dev. 50(3), 43–59 (2002)DE Meyer, RW Schvaneveldt, Facilitation in recognizing pairs of words: evidence of a dependence between retrieval
operations. J. Exp. Psychol. 90, 227–234 (1971)C Mitcham, Thinking Through Technology (The University of Chicago Press, Chicago, 1994)S Ohlsson, Deep Learning: How The Mind Overrides Experience (Cambridge University Press, Cambridge, 2011)H Pashler, M McDaniel, D Rohrer, R Bjork, Learning styles concepts and evidence. Psychol. Sci. Publ. Interest.
9(3), 105–119 (2008)D Schunk, Learning Theories: An Educational Perspective. Learning theories: An educational perspective (6th edit) (Pearson
Education Inc, Boston, 2012)SA Sloman, The empirical case for two systems of reasoning. Psychol. Bull. 119(1), 3 (1996)KE Stanovich, Who is Rational?: Studies of Individual Differences in Reasoning (Lawrence Elrbaum Ass, Mahwah NJ, 1999)KE Stanovich, The Robot's Rebellion: Finding Meaning in the Age of Darwin (University of (Chicago Press, Chicago, 2004)R Sun, SM Lane, RC Mathews, The two systems of learning: An architectural perspective, in Two Minds: Dual Processes
and Beyond, ed. by JSBT Evans, K Frankish (Oxford University Press, Oxford, 2009), pp. 239–264J Sweller, The worked example effect and human cognition. Learn. Instruct. 16(2), 165–169 (2006)A Tversky, D Kahneman, Judgment under uncertainty: heuristics and biases. Science 185(4157), 1124–1131 (1974)LS Vygotsky, Mind in Society: The Development of Higher Psychological Processes (Harvard University Press, Harvard, 1978)
doi:10.1186/s40561-014-0005-4Cite this article as: Koper: Conditions for effective smart learning environments. Smart Learning Environments2014 1:5.